jloyd

I am in a bit of a situation. I am working with a Hurco Autobend 7 controller on a metal press brake (big bending machine) I am writing a program to transfer jobs to and from the controller. I have contacted Hurco and they ahve very little to tell me. The software is a blend of outsoucred code and in house customizing.

Anyway, my problem

The controller gives me a record package in a big byte array. This array contains all of the data for a particular job. Strings are store with there corresponding byte values, integers are store in two bytes, little endian. Floating point numbers are stored in 4 bytes (I assume little endian). I have tried to process the bytes using IEEE 745 type conversion but it fails. Are there any other methods out there?

Example data: 123.45 stored as {0x0E,0xFFFFFF87,0xFFFFFF9B,0x40}

jim mcnamara

Part of the arpanet package in most C installations, there are two functions in netinet/in.h in both windows and unix:

htons, ntohs which are designed to swap endianness for short integers -ie, two byte integers.

As for converting floating point - what is your OS?

jloyd

The target system is a Win98 machine. The data was captured with a Win XP machine

jim mcnamara

Do you have a clue what format the data is in? Microsoft has different FP fornmats -
for example see:

http://support.microsoft.com/default...kb;en-us;11936

DaWei

That doesn't look like it would make a valid floating-point number (well, a non-extreme number), however one arranged it, by any common method of representation. If you can't discover the actual procedure used, you might toss us a few patterns and their known-correct values to diddle with. Include a positive and negative number with the same absolute magnitude.

__________________
Abstraction doesn't make it impossible to write bad code; it makes it possible to write superior code.
Contributor's Corner: Grumpy on C++ Exceptions DaWei on Pointers